EU contract number RII3CT-2003-506395
CARE Conf-05-043-SRF
SRF
BEHAVIOUR OF GAS CONDITIONS DURING VACUUM ARC DISCHARGES
USED FOR DEPOSITION OF THIN FILMS
P. Strzyzewski*, L. Catani●, A. Cianchi¤●, J. Langner*, J.Lorkiewicz¤, R. Mirowski*, R.
Russo¤¶, M. Sadowski*, S. Tazzari¤● and J. Witkowski*
* The Andrzej Soltan Institute for Nuclear Studies, 05-400 Otwock/Swierk, Poland, e-mail:
[email protected]¤ Dipartamento di Fisica, Universita degli studi Roma “Tor
Vergata”, 00133 Roma, Italy ● INFN-Roma2, Via della ricerca scientifica 1, 00133 Roma,
Italy ¶ Instituto di Cibernetica “E. Caianiello”, CNR, 80078 Pozzuoli, Italy
Abstract
The paper concerns an important problem which is connected with the inclusion of some
impurities in the deposited metal film. It was found that appearance of contaminants in the
film is induced mainly by water vapor remnants inside the vacuum chamber. The paper
presents information on changes in the gas composition during and between arc-discharges,
which is of primary importance for the selection of appropriate experimental conditions.
Contribution to the International Conference PLASMA-2005, Opole, Poland
Work supported by the European Community-Research Infrastructure Activity under the FP6
“Structuring the European Research Area” programme (CARE, contract number RII3-CT2003-506395).
1
EU contract number RII3CT-2003-506395
CARE Conf-05-043-SRF
Behaviour Of Gas Conditions During Vacuum Arc
Discharges Used For Deposition Of Thin Films
P. Strzyzewski*, L. Catani●, A. Cianchi¤●, J. Langner*, J.Lorkiewicz¤, R.
Mirowski*, R. Russo¤¶, M. Sadowski*, S. Tazzari¤● and J. Witkowski*
* The Andrzej Soltan Institute for Nuclear Studies, 05-400 Otwock/Swierk, Poland,
¤
e-mail: [email protected]
Dipartamento di Fisica, Universita degli studi Roma “Tor Vergata”, 00133 Roma, Italy
●
INFN-Roma2, Via della ricerca scientifica 1, 00133 Roma, Italy
¶
Instituto di Cibernetica “E. Caianiello”, CNR, 80078 Pozzuoli, Italy
Abstract. The paper concerns an important problem which is connected with the inclusion of some impurities in the
deposited metal film. It was found that appearance of contaminants in the film is induced mainly by water vapor remnants
inside the vacuum chamber. The paper presents information on changes in the gas composition during and between arcdischarges, which is of primary importance for the selection of appropriate experimental conditions.
Keywords: Catodic Arc, Arc Deposition, Superconducting Film, RF Cavity, Quadropole Mass Spectrometer
PACS: 52.77.Dq, 52.80.Mg, 52.80.Vp
INTRODUCTION
The vacuum arc, which is one of the oldest techniques used for the deposition of thin films, is now widely used
for the Plasma Immersion Ion Implantation and Deposition (PIII&D) in laboratory and industry [1]. Despite of high
progress in this field observed during last three decades, involving e.g. magnetic filters for the elimination of microdroplets, some problems have not been resolved so far. One of them is believed to originate from the incorporation
of impurities in a metal film. High adsorption of contaminants by thin films refers mainly to so-called getter
materials, which can absorb impurities from the surrounding and dissolve them inside the deposited layer.
The paper concerns an important problem which is connected with the inclusion of some impurities in the
deposited Nb film. It was found that the appearance of contaminants in the deposited film is induced mainly by
water vapor remnants from the vacuum chamber. Particular attention is paid to a comparison of different gas
conditions during arc discharges at high-vacuum conditions (at the background pressure in the range of 10-8-10-7
hPa) and at UHV experiments (at the background pressure within the range of 10-11-1010 hPa).
NIOBIUM AS A GETTER
Getters are well known and often used in nowadays science and industry. The term “getters” was first introduce
in 1882 by Thomas Edison’s assistant Malignani. He developed a technique of the red-phosphorous deposition for
increasing a lifetime of incandescent lamps. The red phosphorous reacts with water vapor and breaks the watertungsten cycle. This process is still in used in the lamp manufacturing. Moreover, getters are common in electron
tubes and cathode ray tubes for television and computer monitor applications. The getter technology expanded also
into a broad array of technologies and industries, e.g. in vacuum pumps for ultra-high vacuum (UHV) systems and
particle accelerators, as well as in gas-purification systems for semiconductor processes and nanotechnology. Getters
are used in these systems to pump away contaminants and/or keep the constant pressure inside. They work by the
adsorption of gasses and chemical reactions. In a consequence, the layers of non-reactive oxides, carbides and
nitrides can be created depending of adsorbed-gas type.
2
EU contract number RII3CT-2003-506395
CARE Conf-05-043-SRF
Recently the vacuum arc technology was proposed as a possible alternative for depositing thin superconducting
films of pure niobium on the inner surface of RF cavities for particle accelerators. Presence of the impurities usually
deteriorates superconducting properties of niobium films. Since niobium is an excellent getter for vacuum residual
gases a special technique of the thin film deposition must be applied. A new UHV cathodic-arc technique for the
coating copper RF cavities was proposed in 2000 [2-3]. Results of residual gas conditions during deposition process,
which are presented in the paper, show that numbers of contaminants are strongly reduced. Lowering the pressure to
UHV standards can open way towards the production of ultra-pure metallic films. Preliminary studies, as performed
at the IPJ in Swierk, Poland, in the collaboration with the Tor Vergata University team in Rome, Italy, have already
shown that thin niobium films, which were deposited by arc-discharges under UHV conditions, demonstrate similar
properties to the pure bulk niobium [4-5]. The UHV arc-deposited Nb-layers on sapphire substrates have been
characterized by measuring their critical temperature Tc and RRR, defined as the resistivity at room temperature
divided by the resistivity at 10 K. The critical temperature of the deposited material is in general very sensitive to
impurities, e.g. very small amounts of oxygen in the Nb-film can lower its Tc value significantly. The Nb RRR is
also very sensitive to impurities. Typical RRR values for Nb-films deposited by sputtering at a room temperature
have ranged from 20 to 50.
EXPERIMENTAL SET-UP
For the coating of inner surfaces of RF accelerating structures, there were proposed various approaches. First one
has been based on a linear (cylindrical) geometry of the arc source. A cylindrical cathode of the linear-arc can be
placed along the cavity axis. Changing a position of a permanent magnet, which might be located inside of the
tubular Nb cathode, one can drive the arc discharge along the symmetry axis - in order to obtain uniform coating of a
single cavity or multi-cell structure. A scheme of such a system, which has been designed at IPJ, is shown in Fig.1.
The facility is pumped down to 10-11 hPa with an
oil-free pumping system consisting of a two-stage
piston pump as a fore-line of a turbo molecular pump.
The low pressure is achieved only when all parts of
the deposition system are designed and built in
accordance with the UHV-technology requirements.
In our case, all the vacuum chamber components and
accessories, as well as all vacuum connections, have
been manufactured using only high purity materials:
stainless-steel, OFHC copper and high-quality
ceramics (shielded from the arc). The cathode and all
parts accessible to the arc have been made of pure Nb
with RRR ≥ 250.
The reliable triggering (ignition) of arc discharges
is often a serious problem, even in the industrial arcbased devices. In HV systems, thin layers of gasses
and impurities formed upon the surface of electrodes,
are beneficial in that sense that they facilitate the
starting of an arc discharge. Under UHV conditions
the high-temperature baking of the vacuum chamber
removes almost totally such layers. The described
effects, as well as requirements that all other sources
of impurities must be effectively eliminated, make the
arc ignition more difficult. After testing many known
triggering methods from the point of view of the
FIGURE 1. Schematic drawing of an UHV apparatus with
operational reliability and cleanness, we have finally
the linear-arc source for single-cell cavity coating.
decided to use a laser beam focused upon the cathode
surface (through an appropriate vacuum-tight glass
window). Under such conditions the arc discharge can be triggered extremely reliably without introducing any
additional impurity whatsoever [6]. Our arc sources have been equipped with Nd:YAG lasers (50-100mJ, 10ns), and
the mastering of the laser ignition technique has appeared to be decisive for improving properties of the deposited
superconducting films.
3
EU contract number RII3CT-2003-506395
CARE Conf-05-043-SRF
COMPARISON OF HV AND UHV CONDITIONS
The composition of residual gases before and during the coating process is monitored with a Residual Gas
Analyzers (RGA). A Quadropole Mass Spectrometer, used in the experiment, seems to be a very useful tool for
qualitative and quantitative analysis of residual gases at UHV conditions. Typical examples of mass-spectra,
presenting gas conditions at pressures 10-8 hPa (before baking) and 10-10 hPa (after baking), are shown in Fig.2.
FIGURE 2. Typical RGA mass-spectra recorded before and after 24-hour baking at 150 0C
The presented mass-spectra demonstrate very well differences between the HV and UHV conditions. One can see
that reaching UHV standards can cause the practical elimination of impurities from the vacuum chamber, i.e. water
remnants (mass 18), nitrogen (mass 28) and CxHy groups (associated with masses 38-43). Under such conditions
hydrogen became the most dominant gas. Obtaining such “clean” conditions allow us to start the deposition process.
FIGURE 3a. Behavior of the partial pressure (in ion-current
units) residual gases in HV conditions during the arc.
FIGURE 3b. Behavior of the partial pressure (in ion-current
units) residual gases in UHV conditions during the arc.
Figs. 3a and 3b show results of the time-resolved measurements of the selected ion masses at the HV and UHV
conditions. Each of the diagrams presents three pulses of gas during the arc discharge, with some breaks between.
Length of the discharge pulse was about 2-3 minutes. In both cases, immediately after the start of arc discharge, one
could observe the out-gassing of gases from the cathode. The pressure increased up to 10-6 -10-7 hPa when the arc
discharge was started, and it remained almost stable (at the latter value) throughout the whole deposition process.
The gas pressure rise during the arc discharge was found to be almost exclusively caused by hydrogen. Its partial
pressure at the UHV condition is usually about 2-3 orders of magnitude higher than that of other gases. However,
when the deposition is performed at HV, the level of hydrogen was only 1-2 orders of magnitude higher. After 40-60
seconds the partial pressure of out-gassed species, excluding water vapor, began to fall slightly. After the start of the
arc at HV conditions some decrease in water vapor pressure was observed. It was probably caused by the formation
of a fresh oxide-free niobium surface, which could absorb water vapor strongly. Such phenomena have also been
observed during arcing at UHV conditions, where amount of water vapor is the smallest in comparison to H2, CO2
and N2. As soon as the arc extinguishes the hydrogen pressure drops about one order of magnitude, reaching its
equilibrium pressure in presence of the niobium surface, and then starts slowly decreasing, pumped by the 200 l/s
turbo-molecular pump. Water vapor becomes again the most dominant gas in HV conditions. Figure 3b presents also
4
EU contract number RII3CT-2003-506395
CARE Conf-05-043-SRF
behavior of gases when laser triggering pulse hit the cathode but arc discharge doesn’t start. This situation is shown
as a first short peak.
Figures 4a-4d present compositions of residual gases in four different cases and they allow easy to compare the
HV and UHV conditions. The measurements were performed before arc discharges (as the steady-state case) and
about 30 s after the arc stopping. It is important to note that a percentage of water dropped from 80% at HV before
the discharge - to about 3 % at UHV after the discharge. In that case the residual gas (inside the vacuum chamber)
was mainly hydrogen. Percentage of other gases could be neglected.
a
b
c
d
-8
FIGURE 4. Residual gas compositions in vacuum chamber in: HV conditions (“a” – before arc discharge at 5x10 hPa, “b” –
after arc discharge at 9x10-7 hPa) and UHV conditions (“c” – before arc discharge at 6x10-10 hPa, “d” – after arc discharge at
9x10-9 hPa).
SUMMARY
According to our previous papers all impurities, which are present at HV conditions, can considerably influence
quality of the deposited superconducting niobium films. This paper has clearly showed differences between the HV
and UHV conditions. Performing depositions by means of vacuum arc discharges at the UHV conditions can
guarantee a very low level of unwanted gases, mainly water vapor. This is especially important during dc- or pulsedtype arc-discharges, as well as after them, because it prevents oxidizing of the surface layers during the cooling
down process.
ACKNOWLEDGEMENT
We acknowledge the support of the European Community-Research Infrastructure Activity under the FP6
“Structuring the European Research Area” programme (CARE, contract number RII3-CT-2003-506395).
REFERENCES
1.
2.
3.
4.
5.
6.
A. Anders, Vacuum 67, 2002, 673.
J. Langner, in TESLA Report 2000-15. Edit.: D. Proch, 2000 DESY.
R. Russo, et al., in Proc. X Workshop on RF superconductivity, Tsukuba 2001, KEK Proc. 2003-2, 4.
J. Langner, et al., Czech. J. Phys. 54, Suppl.C, 914 (2004).
R. Russo, et al., Supercond. Sci. Technol. 18, 2005, L41-L44.
J. Langner, at al., Proc. International Conference on Plasma Research and Applications PLASMA-2003, Warsaw, Poland,
September 9-12, 2003.
5